GPC3 imaging of the indeterminate liver nodule, led from China
Cirrhotic livers throw up nodules that scans cannot classify, and the usual answer is to wait three to six months to see if they grow. Several Chinese hospital programmes are testing tracers that recognise a protein carried by most liver cancers, and almost none of that work appears in Western pipeline reviews.
Overview
This record exists partly to demonstrate the non-US-registry lens. The clinical problem is well defined: using the CT/MRI LI-RADS algorithm, the pooled proportion of hepatocellular carcinoma is 31% for LR-3 and 64% for LR-4 observations, so a large share of nodules sit in categories that neither confirm nor exclude cancer, and the management default is serial imaging. Glypican-3 is overexpressed in 70 to 80% of hepatocellular carcinomas and virtually absent in normal adult liver and benign hepatic lesions, which is an unusually clean contrast for a liver target.
The programmes divide along geography. The most mature single dataset is Western: [68Ga]Ga-RAYZ-8009 from RayzeBio, now part of Bristol Myers Squibb, with a 24-patient first-in-human case series run in the Netherlands and Germany, no adverse events, and tumour-to-liver ratios averaging 8.3 at about an hour. A related investigator-initiated cohort in India reported 64 patients scanned with 92% showing at least one positive lesion. But the deepest programme is at Wuhan Union Hospital: 68Ga-XH06, an antibody-fragment immunoPET agent registered as NCT06383520 with 100 patients estimated, published as a 5-patient pilot in EJNMMI and an 8-patient first-in-human study in Clinical Cancer Research, with conference cohorts of 30 and 36 patients reporting sensitivity of 90.63% and specificity of 100% against pathology in one data cut and 91.7% and 83.3% in another. A head-to-head against enhanced MRI is registered at 150 patients, and a further 40-patient study is recruiting at Xiamen.
No GPC3 imaging agent has FDA, EMA or NMPA marketing approval. The point of the record is that a landscape review built only on ClinicalTrials.gov Western entries would see the 24-patient case series and miss the registered 100- and 150-patient Chinese studies behind it.
When a nodule in a cirrhotic liver is categorised LR-3 or LR-4, no test says whether it is cancer now, and the default is three to six months of serial imaging in a disease where delay costs the chance of curative treatment.
- The strongest accuracy figures are conference abstracts from overlapping cohorts at one hospital, and the exact patient overlap between the 5, 8, 30 and 36-patient reports is not stated in any source.
- Around 20 to 30% of hepatocellular carcinomas do not overexpress glypican-3, so a negative scan cannot exclude cancer and the test would have to be positioned as a rule-in test only.
- No dosimetry has been published for the newer gallium-68 constructs.
- Results from the registered 150-patient head-to-head against enhanced MRI, which is the study that would establish a clinical claim.
- Whether Bristol Myers Squibb advances RAYZ-8009 beyond a case series, which would signal industry conviction.
- Peer-reviewed publication of the 30 and 36-patient Wuhan cohorts, which currently exist only as conference abstracts.
The five domains of a radiopharmaceutical due-diligence review, adapted from the Fryback-Thornbury hierarchy of diagnostic efficacy. Each carries the state of the public evidence and the fact behind it, not a score out of five: a single total hides the reason and stops telling candidates apart. Domains marked as carrying most weight are the ones that decide the case at this asset's stage.
Glypican-3 is a cell-surface heparan sulfate proteoglycan overexpressed in 70 to 80% of hepatocellular carcinomas yet virtually absent in normal adult liver and benign hepatic lesions.
A 24-patient first-in-human series, a 64-patient investigator-initiated cohort in India, a peer-reviewed 5-patient and 8-patient Chinese series, and registered Chinese studies of 100 and 150 patients; none is registrational.
Tumour-to-liver ratio averaged 8.3 at about one hour in the first-in-human series; conference cohorts report sensitivity 90.63% with specificity 100% in 36 patients and 91.7% with 83.3% in 30, both against pathology and both unpublished in full.
No adverse events reported across the published cohorts, but no dosimetry in mSv or cGy/MBq was found for any of the newer gallium-68 GPC3 probes.
No GPC3-targeted imaging agent has received FDA, EMA or NMPA marketing approval, and no filing has been identified in any jurisdiction.
Red, amber or green on each review category, each with its reason, so an obvious no-go takes seconds rather than a meeting.
The Chinese lead is an antibody-fragment immunoPET agent and the Western lead a synthetic peptide; both image within about an hour, but liver background is the hardest part of this anatomy and the constructs differ in kinetics.
31% of LR-3 and 64% of LR-4 observations prove to be cancer, and the alternative to a decisive test is months of watching.
The peptide lead sits inside Bristol Myers Squibb after the RayzeBio acquisition and the antibody-fragment lead with a Chinese academic hospital; several other constructs are in early trials.
Standalone: the scan resolves a management dilemma by itself, in a disease that is a leading cause of cancer death worldwide, with the largest patient population in China.
Gallium-68 across all the leading agents.
At least five separate programmes have human data, split between Western peptide and Chinese antibody-fragment approaches, and none is close to filing.
- RayzeBio (Bristol Myers Squibb)first-in-human
[68Ga]Ga-RAYZ-8009, 24-patient first-in-human series in the Netherlands and Germany, no adverse events, mean tumour-to-liver ratio 8.3.
https://pubmed.ncbi.nlm.nih.gov/39362766/ - Union Hospital, Tongji Medical College (Wuhan)early phase 1, recruiting
68Ga-XH06 immunoPET: NCT06383520 at 100 patients estimated, a registered 150-patient head-to-head against enhanced MRI, and published pilots of 5 and 8 patients.
https://clinicaltrials.gov/study/NCT06383520 - All India Institute of Medical Sciencesinvestigator-initiated
[68Ga]Ga-RYZ-GPC3 in 64 patients, with 92% showing at least one positive lesion.
https://aacrjournals.org/cancerres/article/84/6_Supplement/2585/740458/ - First Affiliated Hospital of Xiamen Universityrecruiting
68Ga-GPC3 PET/CT against contrast-enhanced MRI, 40 patients estimated, recruiting.
https://clinicaltrials.gov/study/NCT07677774
Similar pages
not linked directly; found by shared links- Target5T4 (trophoblast glycoprotein)
Shares Glypican-3, Immuno-PET, Gallium-68 and the tag China.
- OpportunityDLL3 imaging for small-cell lung cancer
Shares Immuno-PET, PET/CT, Gallium-68.
- OpportunityTROP2 imaging where no tissue test is required
Shares Immuno-PET, PET/CT, Gallium-68.
- PersonAndrei Iagaru
Shares PET/MRI, Gallium-68.
- RoadmapMolecular imaging roadmap: FDG → PSMA → FAP → antigen and immune PET
Shares PET/MRI, Immuno-PET, PET/CT, Gallium-68.
- OpportunityClaudin-18.2 imaging when one biopsy says negative
Shares Immuno-PET, PET/CT, Gallium-68.
- InstitutionChinese PLA General Hospital
Shares Hepatocellular carcinoma, PET/CT.
- OpportunityCEACAM5 imaging for occult colorectal recurrence
Shares Immuno-PET, PET/CT, Gallium-68.